This application relates generally to image processing. More specifically, the application relates to identifying image orientation. Yet more specifically, the application relates to identifying image orientation in medical and dental X-ray shadow-grams.
Images produced in conventional roll-film cameras, on film, are easy to orient correctly because the camera is constructed in such a way that the emulsion of the film always faces the lens. Because of the shapes of film cassettes and cameras, the film cannot be inserted in modern cameras with the emulsion facing away from the lens, so it is always known that the light, or other radiation recorded by the film, struck the film from the emulsion side. Thus, when orienting slides for projection, or film images for viewing on a light box, it is always known to put the emulsion toward the projection lens in the slide projector, or to put the film on the light box with the emulsion towards the viewer. When that is done, the image seen by the viewer, whether projected or viewed on the light box, will have a known defined correspondence with the orientation of the objects in the original scene. Moreover, even if orientation is lost, reorienting based on determining where the emulsion is will reestablish proper orientation.
Likewise, digital images produced using cameras with detectors such as charge-coupled devices (CCDs) that are sensitive to exposure from one side only are inherently unambiguous. The cameras used to produce such images are physically constructed and arranged with a lens or lens mount in a fixed position relative to the detector, so that the direction from which the exposing radiation strikes the detector is always known. CCD sensors used in dental digital radiography also are physically constructed and arranged to produce a diagnostic image only when exposed from the proper direction. Such CCD sensors include a radiopaque element on the side away from the intended direction of exposure.
The problem is somewhat more complicated for dental or medical diagnostic images produced on conventional film. One common type of dental or medical diagnostic image is a shadow-gram produced by placing a radiographic film on one side of the object to be imaged and a source of radiation to which the film is sensitive on a diametrically opposite side of the object to be imaged. The radiation, for example X-radiation, casts a shadow of the object to be imaged on the film, thus revealing density variations within the object whose shadow has been cast. When a standard orientation of the examined object is used, and the film is viewed from the same side as the radiation source exposing the emulsion, the laterality, i.e. right versus left, of the object is preserved, because the viewer can unambiguously discern right from left side by observing the image alone. However, medical or dental X-ray film can be exposed and viewed from either side because the film is transparent to both X-rays and visible light. To further complicate matters of orientation, an emulsion is often applied to both sides of the film in order to lower the dosage of radiation required to produce the diagnostic image, by increasing the sensitivity of the film. However, film thus constructed cannot be oriented on the basis of which side has an emulsion. Thus, it is desirable to identify from which side the film was exposed to radiation and so indicate the side from which it should be viewed.
Conventionally, and due in part to its high sensitivity to visible light, as well as to X-rays, radiographic film is usually held in an opaque cassette which not only prevents exposure to visible light, but also limits X-ray exposure to one side of the film, only. The side of the cassette through which large format film, such as that used for medical X-ray applications, is exposed often includes radiopaque labels indicating various patient information, including patient name, date of exposure, side of the patient (e.g., left or right arm), etc., for example, which provide a clear indication embedded in the image of the side of the film from which the exposure was made. If such labels are used, reversal of the image by viewing from the side opposite the side of exposing incident rays also causes the label characters to reverse, clearly indicating the reversed orientation of the diagnostic image. Systems also exist (e.g., Planmeca's pantomograph) which incorporate within the film cassette a mechanism which automatically optically imprints on the margin of the film, and therefore the image, pertinent patient, exposure, and orientation data during the exposure of the film. Often physical features, e.g., notches in the edge or corner of the film or keyways that orient the cassette itself to a cassette holder within the apparatus exposing the object, are used to further ensure that a consistent orientation of the film in the cassette and the exposure apparatus, and thus, the relative orientation of film and the image recorded on it to the object, is obtained and retrievable independently of the structure represented in the diagnostic image itself. The “back side” of the cassette is itself at least partially radiopaque over its entire surface, so as to prevent accidental exposure from the wrong side. The spatial orientation of the original form creating the shadow in the image is thus clearly and unambiguously defined. However, this clear and unambiguous outcome relies either upon the mechanical features noted above, or upon the X-ray technician properly loading the cassette and placing the orienting labels, followed by exposing from the correct side. Otherwise, if the positioning of the film or the cassette is improper, the image produced visually reflects the improper orientation and the process and must be repeated.
Unlike the large format films described above, dental intraoral films are conventionally provided prepackaged in packets, which may be disposable and which are often flexible. Because intraoral films must be small enough to be positioned within the mouth, there is little room available for the types of labels described above. In the case of conventional dental intraoral films, proper orientation is identified by a raised or embossed bump on the film and packet that points away from a radiopaque backing in the packet that prevents or limits inadvertent exposure from the wrong side. The bump, which projects from the plane of the film in the direction from which the shadow is cast, is a permanent feature of the film and persists as marker of the orientation in the dimension perpendicular to the plane of the film, providing a method by which the viewer can identify the side of exposure. If an exposure is made from the wrong side, the resulting image is recognizably undiagnostic because of its degradation caused by the radiopaque backing contained within the packet. The radiopaque backing, typically a textured heavy metal foil, causes exposure from the wrong side to appear in the image as a textured pattern, while the embossed feature preserves and identifies the orientation information for the viewer. As with medical X-rays in which a proper outcome is assured by correct loading of the cassette, a proper outcome using intraoral dental films is assured only by correct assembly of the packet at the factory, followed by correct handling by the dental X-ray technician. The degradation of the image evidencing improper exposure orientation of the film, which results in the reversal of the diagnostic image when the bump feature is used to orient the image, is, however, immediately obvious to the viewer, assuring that mistaken identification of features in the image is highly unlikely and traceable. This combination of an orientation feature and a feature preventing exposure from the wrong side is now conventional.
By convention, dental intraoral films are produced in a variety of generally rectangular standard sizes, with rounded corners. Also by convention, when prepared for viewing, they are grouped in an anatomical arrangement in a holder called a mount. Standard mounts hold films in one of two orientations only: with the longer dimension of the film in a horizontal orientation, henceforth in this application referred to as “landscape” orientation; and with the longer dimension of the film oriented vertically, henceforth referred to in this application as “portrait” orientation.
In keeping with conventions of viewing of radiographic images, on film or on a computer monitor, and for nonambiguity and clarity of the descriptions, the frame of reference for principal directions is based on the image plane, itself. Up shall be a direction generally from any point in the image toward a top of the edge of the image, while down is in an opposite direction. With the up direction oriented in a natural fashion for a viewer, left and right correspond to the viewer's left and right. Vertical corresponds to the direction of a line running up and down from any point in the image.
U.S. Pat. No. 4,625,325 describes a method which incorporates both conventional radiographic film and phosphor in the same process with radiopaque material. The device described patented therein includes a film packet, similar to the intraoral dental film packet described above. The device also incorporates a pocket for holding a plate which is inserted by the technician prior to the exposure of the film. The plate, which is coated with a phosphor on the side facing the film and radiation source and which incorporates radiopaque material on the opposite side, is positioned adjacent to the film in the packet. The plate is an image amplifier. During the exposure, the film is exposed both by X-rays impinging directly on the film and by the phosphorescence emitted by the plate phosphor stimulated by the same X-rays. Those X-rays which have passed both the film and the phosphor are absorbed by the radiopaque backing, thereby limiting tissue exposure downstream of the recording surfaces. The packet, film and plate are held by a jig in a position such that the film and plate can only be exposed from one side. The phosphor only functions as an amplifier to provide improved signal-to-noise ratio and to lower the radiation dose per exposure. The phosphor film does not function as a storage phosphor holding a latent image, and is not scanned to produce a diagnostic digital image. Furthermore, no issues of image orientation ambiguity from a digital electronic image result from the process, because the image is recorded on conventional radiographic film, as before.
Radiosensitive storage phosphor plates have recently started displacing conventional radiographic emulsion film for recording medical and dental images. Advantages including superior sensitivity, lack of dependence on toxic chemical processing fluids, relative insensitivity to ambient light, reusability, and the ease of digital data storage and transmission all stimulate the growth of this technology. The image orientation issues noted above with respect to conventional film; however, as well as new image orientation problems, for example resulting from the use of image processing software, manifest themselves in the use of such radiosensitive plates. As discussed above, the orientation of the image produced by film technology was identified unambiguously by the presence of a three-dimensional object, namely the bump protruding from the film surface toward the object casting the shadow, providing a permanent and absolute reference in a dimension perpendicular to the plane of the film. The other two dimensions, superior-inferior and anterior-posterior are inferred from the anatomical structures in the image. Unlike the image incorporated into a three-dimensional physical object, i.e., the film with a bump, the images produced by existing phosphor plate technologies are stored and displayed as two-dimensional views without a complete, definite, and permanent indicator of the direction of exposure or of viewing. Incomplete references or markers, as discussed below are incorporated into the existing systems; however, none of the systems are unambiguous, permanent, or complete by virtue of their design.
Although radiosensitive plates are sensitive to a specific, diagnostic radiation type, e.g., X-rays, they are substantially insensitive to visible light for the purposes of registering an image. They can be handled in ordinary room light, absent the usual cassette until the time of the exposure. They are, for sanitary and other purposes such as reducing wear and tear on the phosphor, inserted into radiolucent plastic film sleeves before each use. The plates are also reusable to produce multiple images over a period of time. Erasure, by prolonged exposure to intense visible light, and repackaging in the disposable radiolucent plastic film sleeves is done by a technician at the point of use, rather than at the point of manufacture.
A scanner, through laser illumination, stimulates the phosphor to emit light in an amount which depends on the amount of prior exposure to X-rays, and which in turn is registered as data signal. Currently, commercially available phosphor-based digital radiology systems use plates having a polymer sheet substrate supporting a pastel-colored phosphor layer applied to one surface of the substrate. The other surface of the substrate appears black. For the purpose of this application, the side of the plate which is intended by the manufacturer as the preferred side to be read, e.g., by the scanner, to produce the diagnostic image shall be referred to as the “front side”, while the opposite side of the plate and sensitive layer shall be referred to as the “back side” henceforth. For practical reasons related to the current technology, the side of the plate that is generally intended as the side to be scanned (“front side”) is also the side on which the sensitive layer is nearer the surface of the plate and is visible, thereby available for excitation by the scanning mechanism.
Even though the sensitive layer, e.g., phosphor, is available for scanning from one side, the “front side”, only, it can be exposed and register a latent image from either side, and in some commercial systems equally well. As a result of the possibility that the shadow recorded by the sensitive layer, e.g., phosphor, could have been cast from either side of the plane of the plate, the recorded latent image as well as the visible image resulting from its scan are ambiguous with respect to their laterality. Thus, when recording bilaterally symmetric structures, e.g., left or right jaw, mirror images result, which can be easily confused since they are not uniquely oriented. Therefore, the two sides of the body (or mouth) can be confused by the viewer of the image, resulting in erroneous diagnosis and/or treatment. As a result, current commercially available phosphor plate systems, such as those produced by Air Techniques Inc. and Gendex™, provide detailed explicit instructions to package and expose plates in a specific orientation, so as to preserve image orientation.
Some digital systems employ techniques analogous with technology used in conventional emulsion film. For example, Digora™ (available from Soredex) system phosphor plates incorporate a slightly radiopaque layer on the “back side” of the plate, which reduces patient irradiation by rays that pass through the plate and into the patient. The radiopacity of the backing is featureless, but the backing degrades an image exposed by irradiation from the “back side.”
Systems such as Scan-X™ (available from Air Techniques Inc.) and Denoptics™ (available from Gendex Dentsply) each feature a distinctive marker. The markers are distinctively shaped, either a lower case letter “a” opaquely printed over the “front side” of the phosphor or a small open circle evident as an absence of the phosphor in a localized area, respectively. These markers are incorporated at the time of fabrication of the plates by the manufacturer and are constructed in such a way as to always be read from the phosphor by the scanner in a constant fashion independent of any exposure variables. The result of the presence of a marker produced by either of the above variants, i.e. Scan-X and Denoptics, in fabrication is a diminution, or absence, of phosphorescence from the area of the plate so altered, during scanning. This relative lack of signal is reflected in the visible scan image as a distinct shape corresponding to the shape and the location of the marker on the phosphor plate.
Any features of such a marker that are either asymmetric or placed asymmetrically with respect to a vertical axis of symmetry of the plate, or both, become represented by similarly asymmetric features in the image of the plate following a scan. Furthermore, this marker image becomes reversed with respect to its laterality either in respect to its asymmetric location or its internal asymmetry as a result of software horizontal reflection of the scanned image. As the placement of the source of the radiation during the exposure has no influence over the appearance of the image of such marker, this marker is well suited as an indicator of any reversal of laterality of the entire image following the completion of the scan.
For practical reasons, namely because the image produced by each of the commercially available phosphor plate technology systems can be displayed and viewed in one of four orientations only, which orientations correspond to the conventional orientations for emulsion-based radiographic film mounts, and which orientations are separated by steps of ninety degree rotation relative one another, the rotational transformations of the image (and the plate position) in this application will also be confined to ninety degree steps, or multiples thereof. Thus the available set of orientations to consider for any image will be the two “landscape” and two “portrait” possibilities, i.e., one right side up and one up side down for each category.
The configuration shown in FIG. 1 illustrates generally how a dental X-ray plate can be used to produce images, X-ray shadowgrams, of a patient's lower jaw and teeth. The particular physiological structure is illustrative of the symmetry problem discussed above. However, the problem occurs in connection with many physiological structures as a result of the inherent symmetry present in most biological systems, particularly humans and animals. The configuration of FIG. 1 is now described as it relates to conventional digital dental X-ray plates. Later, in the DETAILED DESCRIPTION, the configuration of FIG. 1 is referenced as it relates to aspects of the invention, which may be practiced in this configuration, as well as other suitable configurations.
The anatomical structure illustrated in FIG. 1 is a human lower jaw, i.e., a human mandible 101, having a set of teeth 102 set therein. Although not shown, for clarity, it may be assumed that the mandible 101 and teeth 102 are part of a living patient's body, covered with the soft tissues, etc. This particular patient has a diagnostically significant condition, e.g., an abscess, denoted by filled circle 103. A dental technician, dentist or other has placed a conventional digital dental X-ray plate 104 in the patient's mouth in a position suitable for capturing a shadowgram of three of the patient's posterior teeth 105 on either side of the patient's mouth. Although in actual practice, the plate 104 would be placed close to the teeth whose shadowgram is being recorded, so as to produce a clear image, for the convenience of illustration the plate 104 is shown centered between the teeth of the left side and the teeth of the right side of the patient's mouth. Finally, two alternative locations for X-ray sources, SOURCE L and SOURCE R, are shown. X-ray source, SOURCE R, produces the images shown in FIGS. 2 and 3, while X-ray source, SOURCE L, produces the images shown in FIGS. 4 and 5. FIGS. 2, 3, 4 and 5 are now described, with reference back to FIG. 1, as required.
FIGS. 2, 3, 4 and 5 illustrate four images that can be produced using a conventional digital X-ray plate exposed from each of two sides, using each of two differently located sources, e.g., positioned at SOURCE L and SOURCE R. Because the reference conventional commercial plate has an open circle mark on the “front side” of the plate, in one corner, an open circle 201 appears in each of the images. Images produced by source, SOURCE L, include images 205 of posterior teeth 105 and images produced by source, SOURCE R, include both images 205 of posterior teeth 105 and an image 203 of diagnostically significant condition 103. In conventional digital X-ray plate usage, the open circle is present on the “front side” of the plate and is intended to direct the technician, dentist or other to face that side of the plate toward the exposure source. However, such a consistent usage is not guaranteed.
If the plate were to be consistently exposed from only one side (e.g., “front side”), then this marker would provide an absolute reference of laterality by eliminating confusion introduced by horizontally flipping of the image. Such reflection results in the displacement of the open circle marker from the lower right or the upper left corners of the image to the lower left or the upper right in the images which are in “landscape” orientation. The displacement would be the reverse for “portrait” orientation. Although recommended, and consistent with best practice, such a consistent exposure of the plate from the sensitive side is not guaranteed either in the loading of the plate into the sleeve or in its placement during exposure itself.
It should be noted that certain predictable rules of translocation govern a system composed of a radiographic plate, the image it holds, the long and the short axes of symmetry of the plate or its image, and a universe of combinations of two motions, a reflection through a vertical plane perpendicular to that of the image and a 90 degree rotation around an axis perpendicular to the plane of the plate at the intersection of its long and short axes of symmetry.
The reflection of the image within this system can occur through one of two modes. The first mode involves casting the shadow, i.e., registering the image, onto the sensitive layer, e.g., phosphor, from one side of the plate, e.g., “back side”, and reading it off the opposite aspect of the sensitive layer, e.g., “front side”. Only one instance of this mode may occur per image. This mode of reflection shall henceforth be referred to as “pre-exposure” reflection in this application. The second mode involves the use of an image processing software tool which reflects right-for-left any selected image. The number of instances of this mode of reflection is not theoretically limited. This mode of reflection shall henceforth be referred to as “software” reflection in this application.
Rotation has two distinct modes. The first mode is a physical rotation of the plate, together with its markers, relative to the object to be imaged prior to exposure involved in changing the orientation from “landscape” to “portrait”, and if continued, back to “landscape”. This mode of rotation shall be referred to henceforth in this application as the “pre-exposure” mode of rotation. The second mode of rotation can occur several ways. After exposure, the plates, being small, unattached objects are free to be moved and become randomized in orientation. These plates are later removed from their sleeves and arbitrarily rotated as to fit into a plate holder mechanism of the scanner, the constraint at this stage being that the “front side” must face the sensor. Once the images are produced on the computer screen, the operator uses the image processor to align the images in proper superior-inferior orientation by rotating them. This software-mediated rotation is limited to multiples of ninety degrees and is the mechanism through which correct “landscape” and “portrait” orientation as well as superior-inferior relationship is achieved as needed. All three of the rotations given above maintain the relationship between the location of the marker and the details of the shadow-gram. They also preserve laterality. The several mechanisms of rotation comprising the second mode of rotation occur after the sensitive layer is exposed and will henceforth in this application be referred to as “post-exposure” mode of rotation. When a “post-exposure” rotation results in a 180 degree rotation, the resulting transformation is equivalent to a reflection of the entire recorded image through a point located at the intersection of the long and the short axes of symmetry of the plate.
In order to better understand the discussion below in the DETAILED DESCRIPTION of how the structures according to aspects of the invention unambiguously identify correct and incorrect image orientation, first a discussion of possible transpositions of the image is given.
Within the context of this application, the two modes of reflection and two modes of rotation comprise the universe of orientation transformations allowed by the laws of physics and by the graphic functions included in the software of the image processors typically provided with digital radiography systems. For the purposes of this example, and in order to demonstrate the inability of only a conventional marker to differentiate various transformations, a marker 3811, as defined earlier, shall be placed in the lower right corner, as viewed from the “front side”, of a plate in a “landscape” orientation. In FIG. 38 the images 3801, 3802, 3803, 3804, 3805, 3806, 3807 and 3808 within the same row (e.g., 3801, 3802, 3803 and 3804; or 3805, 3806, 3807 and 3808) are related to their neighboring images by one ninety-degree “post-exposure” rotation for each step. On the other hand images within the same column (e.g., 3801 and 3805, etc.) are related by a “software” reflection through a vertical plane perpendicular to that of the plate. As the images 3801, 3802, 3803 and 3804 and also the images 3805, 3806, 3807 and 3808 illustrate, rotation alone allows only two marker 3811 locations each for both the “portrait” and the “landscape” orientations of the image. That is to say that if the image is only rotated (i.e. they lie within the same row in FIG. 38), there are two allowed locations of the marker 3811 for the “landscape” orientation 3801 and 3803 or, the lower right and the upper left corner, and two allowed locations for the marker in the “portrait” orientation 3802 and 3804: lower left and upper right corner of the image. It is also evident from FIG. 38 that for both the “portrait” and for the “landscape” orientation a combination of rotation within the plane and reflection through a perpendicular plane manipulation, as described earlier, is sufficient to generate all the possible locations of the marker. It is also evident that all paths involving an odd number of reflections (i.e. one which generates a net shift of the image to a different row in FIG. 38) are qualitatively different from the paths involving an even number of reflections (i.e. no net shift of row generated in the process). Therefore, assuming an original position of the marker 3811 in the lower right corner in a “landscape” orientation as in image 3801, it is clear that relative to original image 3801, the laterality of image 3803 has not been reversed but that of image 3809 has been. Furthermore, without the knowledge of the original image 3809, 3810 it is possible to deduce its laterality by observation of the marker location and the “portrait” vs. “landscape” orientation of the plate. It should also be noted that teeth of the lower and the upper jaw are significantly enough different to allow their recognition in radiographic images thereby preserving orientation in the superior-inferior dimension. By analogy, if the long arrow 3810 is assumed to serve as a recognizable index for the superior direction, and the short arrow 3809, the forward direction, then only images 3801 and 3805 represent images properly oriented with respect to the superior-inferior dimension. Furthermore, of those two, only image 3801 has preserved the original laterality of the object casting the shadow. Further complicating the situation is that the plate could have been exposed from the “back side” and viewed from the “front side”, resulting in image 3805. This image must be reflected horizontally to be viewed in the “correct” orientation of image 3801.
Suppose, for the purpose of analyzing the images in FIGS. 2, 3, 4 and 5, that the technician, dentist or other exposing this patient's digital dental X-ray plate 104 has oriented the “front side” defined above, carrying the open circle, and consequently the sensitive side of the plate 104, toward the X-ray source located at the position SOURCE L, and in the lower right corner of the plate 104 as viewed from the direction of the X-ray source located at the position SOURCE L. FIG. 3 is the image read from the “front side” of the plate 104, when the plate so oriented is exposed by an X-ray source at the location designated SOURCE R. If the dentist reading such an image is aware that the exposure has been made from the “wrong” side, i.e., the “back side”, of the plate, the dentist can use image processing software to reorient the image by horizontally flipping the image, as shown in FIG. 2. Mark 201 is transposed from the right to the left, side of the image. The same plate 104, oriented the same way, but exposed from an X-ray source at the location designated SOURCE L, when read from the sensitive side of the plate 104, produces the image shown in FIG. 5. The image of FIG. 4 can be inadvertently produced by manipulation of the image processing software to horizontally flip the image of FIG. 4. Since FIGS. 2 and 4, and FIGS. 3 and 5, are respectively indistinguishable without knowing from which side the plate 104 was exposed, there is no way to determine which side of the patient's jaw the condition 103, seen only in the images in FIGS. 2 and 3, is on. If condition 103 does not produce any externally observable symptoms, the X-ray image may be the only evidence upon which the dentist can rely for determining the location to treat. An ambiguity is introduced into the record that cannot be resolved without another exposure of the patient to radiation. Digital plate technology in its current form does not assure that the orientation of the image can be ascertained. Current technology instead relies on the statistical likelihood that the technician will expose the film correctly vast majority of the time. However, no unambiguous marker of the exposure orientation exists within the image. The following four examples illustrate the problem of a lack of an internal reference:
1. If only one image is available for viewing independently of other patient information, the viewer will not be able to identify the correct orientation of the image, except by making an assumption that it was exposed from the “front side”.
2. If a technician is consistently making the error of exposing the films from the “back side”, unknown to the viewer, the viewer will conclude when comparing images that images (in fact) exposed and oriented correctly are incorrectly oriented (which is not factually correct), thus compounding the problem.
3. A disgruntled or incompetent employee can wreak havoc with the records without anyone realizing it, or having a way of tracing the problem by using software to alter the apparent orientation of images in the records.
4. A person with fraudulent intent can expose the plate intentionally from the “back side” in order to make the image appear as though it depicts the opposite side of the body.
The analysis of FIGS. 6, 7, 8 and 9 is similar to FIGS. 2, 3, 4 and 5, respectively, except for the initial orientation of the plate. These images are produced by a plate 104 oriented with an open circle, and consequently the sensitive side of the plate 104 oriented toward the X-ray source located at the position SOURCE L, but in the top left corner of the plate as viewed from the direction of the X-ray source located at the position SOURCE L.
FIGS. 10, 11, 12 and 13 represent images produced by a plate 104 oriented with an open circle, and consequently the sensitive side of the plate 104 oriented toward the X-ray source located at the position SOURCE R, but in the bottom right corner of the plate as viewed from the direction of the X-ray source located at the position SOURCE R.
FIG. 10 shows the image produced by SOURCE R, in which the open circle 201 is in the lower right corner. FIG. 11 can be inadvertently produced by a horizontal flip of the image of FIG. 10 using image processing software. FIG. 12 shows the result of exposing the plate using SOURCE L. In order to view the image in an orientation expected by the dentist, the image of FIG. 12 can be horizontally flipped to produce the image of FIG. 13. Similarly to the situation described above in connection with FIGS. 2, 3, 4 and 5, FIGS. 10 and 12 are inherently indistinguishable, as are FIGS. 11 and 13.
The analysis of FIGS. 14, 15, 16 and 17 is similar to FIGS. 10, 11, 12 and 13, respectively, except for the initial orientation of the plate. These images are produced by a plate 104 oriented with the open circle, and consequently the sensitive side of the plate 104 oriented toward the X-ray source located at the position SOURCE R, but in the top left corner of the plate as viewed from the direction of the X-ray source located at the position SOURCE R.